[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <ba8aa36a-d0d7-b716-a9c7-02c6d5a60712@bytedance.com>
Date: Tue, 15 Nov 2022 11:39:55 +0800
From: Abel Wu <wuyun.abel@...edance.com>
To: Miklos Szeredi <miklos@...redi.hu>
Cc: linux-fsdevel@...r.kernel.org, Dave Chinner <david@...morbit.com>,
Theodore Ts'o <tytso@....edu>, Karel Zak <kzak@...hat.com>,
Greg KH <gregkh@...uxfoundation.org>,
Christian Brauner <brauner@...nel.org>,
linux-kernel@...r.kernel.org,
Linux API <linux-api@...r.kernel.org>,
linux-man <linux-man@...r.kernel.org>,
LSM <linux-security-module@...r.kernel.org>,
Ian Kent <raven@...maw.net>,
David Howells <dhowells@...hat.com>,
Linus Torvalds <torvalds@...ux-foundation.org>,
Al Viro <viro@...iv.linux.org.uk>,
Christian Brauner <christian@...uner.io>,
Amir Goldstein <amir73il@...il.com>,
James Bottomley <James.Bottomley@...senpartnership.com>,
chenying.kernel@...edance.com
Subject: Re: [RFC PATCH] getting misc stats/attributes via xattr API
On 11/14/22 8:35 PM, Miklos Szeredi wrote:
> On Mon, 14 Nov 2022 at 10:00, Abel Wu <wuyun.abel@...edance.com> wrote:
>>
>> Hi Miklos and anyone interested in this proposal, is there any update on
>> this? Sorry that I didn't find any..
>
> No update.
>
> Which part are you interested in?
We noticed that atop(1) can introduce a burst cpu usage once number of
processes becoming large. It is mostly due to the overhead of massive
syscalls. There are similar cases like monitor agents recording system
status and consuming resources in modern data centers. So it would be
nice to get a bunch of info in one syscall.
>
> Getting mount attributes? Or a generic key-value retrieval and
> storage interface?
The latter.
>
> For the first one there are multiple proposals, one of them is adding
> a new system call using binary structs. The fsinfo(2) syscall was
> deemed overdesigned and rejected. Something simpler would probably be
> fairly uncontroversial.
>
> As for the other proposal it seems like some people would prefer a set
> of new syscalls, while some others would like to reuse the xattr
> syscalls. No agreement seems to have been reached.
So the divergence comes from 'how' rather than 'why', right?
Thanks & Best,
Abel
>
> Also I think a notification system for mount related events is also a
> much needed component. I've tried to explore using the fsnotify
> framework for this, but the code is pretty convoluted and I couldn't
> get prototype working.
>
> Thanks,
> Miklos
Powered by blists - more mailing lists